From: Dennis J. Hartmann (dennisjhartmann@hotmail.com)
Date: Thu Apr 28 2005 - 12:27:52 GMT-3
There's a LOT of confusion over this issue.
Here's the difference between Frame-Relay Traffic Shaping (FRTS) and
class-based shaping (shape command used in the modular qos cli [MQC]):
FRTS can do FRF.12 (VoIP over FRAME and CB-shaping can NOT!!! --->
unless you're running latest and greatest 12.3 T code train).
Class-based shaping CAN adapt to BECN messages and they can ADAPT
FECN messages to the source of the traffic by sending Q.922 FRAME test
frames back to the source (so both sides can throttle their rate).
PLEASE..... NEVER do this in the real world! Adaptive traffic
shaping is a BAD thing. Service provider NNI equipment is running at speeed
of OC-12 (622 Mbps) and above (up to OC-192 ~10Gbps). Their TX and RX
buffers flush VERY fast and you will not receive ANY benefit by throttling
your rate (except for punishing yourself). I have personally heard this
from AT&T and MCI engineers.... And I agree with them. But the CCIE lab is
a different topic, now isn't it.
The configurations that Erick Bergquist posted are not identical and
here's why:
Even though we configured a Bc for the shape adaptive command, we do
not use it. Look at my configurations below and notice that regardless of
whether I set a Bc or a Bc + Be, when I shape to AVERAGE, the result is
always the same (the target and average rate are THE SAME!). But when I
shape to peak (class2), the target rate is DOUBLE the average rate because
I'm allowing bursting. The default setting for the target rate is CIR * 1 +
(Be + Bc). Since the defaults for Bc and Be are both 8000 bytes, 8000/8000
= 1 and 1 + 1 = 2. CIR (768kbps * 2 = 1.536 Mbps [1536000 bps]).
If someone has access to a Smartbits 600 packet generator, I'd love
to see the results of this test.....
WARNING!!!!!! Older versions of IOS would use CIR + Bc when shaping
to average..... 12.2(15T) is what I tested with.
R1#show class-map
Class Map match-all test2 (id 2)
Match none
Class Map match-all test1 (id 1)
Match none
Class Map match-any class-default (id 0)
Match any
R1#show policy-map
Policy Map test
Class test1
Traffic Shaping
Average Rate Traffic Shaping
CIR 768000 (bps) Max. Buffers Limit 1000 (Packets)
Bc 7680 Be 7680
Class test2
Traffic Shaping
Peak Rate Traffic Shaping
CIR 768000 (bps) Max. Buffers Limit 1000 (Packets)
Class class-default
Traffic Shaping
Average Rate Traffic Shaping
CIR 768000 (bps) Max. Buffers Limit 1000 (Packets)
Bc 7680 Be 0
Adapt to 8000 (bps)
Map FECN to BECN
R1#show policy-map int s 0/0
Serial0/0
Service-policy output: test
Class-map: test1 (match-all)
0 packets, 0 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: none
Traffic Shaping
Target/Average Byte Sustain Excess Interval Increment
Rate Limit bits/int bits/int (ms) (bytes)
768000/768000 1920 7680 7680 10 960
Adapt Queue Packets Bytes Packets Bytes Shaping
Active Depth Delayed Delayed Active
- 0 0 0 0 0 no
Class-map: test2 (match-all)
0 packets, 0 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: none
Traffic Shaping
Target/Average Byte Sustain Excess Interval Increment
Rate Limit bits/int bits/int (ms) (bytes)
1536000/768000 4800 19200 19200 25 4800
Adapt Queue Packets Bytes Packets Bytes Shaping
Active Depth Delayed Delayed Active
- 0 0 0 0 0 no
Class-map: class-default (match-any)
84 packets, 8133 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: any
Traffic Shaping
Target/Average Byte Sustain Excess Interval Increment
Rate Limit bits/int bits/int (ms) (bytes)
768000/768000 960 7680 0 10 960
Adapt Queue Packets Bytes Packets Bytes Shaping
Active Depth Delayed Delayed Active
- 0 53 8436 0 0 no
R1#
This archive was generated by hypermail 2.1.4 : Tue May 03 2005 - 07:55:10 GMT-3